Hacker News Reader

vitorsron ML system design: 300 use cases from 80 companies

I also highly recommend going through KDD Applied Data Science track papers.

Here’s a Google Scholar search query to get you started:

https://scholar.google.com/scholar?q=%22Applied+Data+Science...

vitorsron Typst – Compose Papers Faster

> Say goodbye to fidgety formula editors and antiquated syntax.

They should not have deviated from the American Mathematical Society notation [1].

Compare the list of supported commands on MathJax [2] or KaTeX [3] and Typst [4].

The main purpose of LaTeX today is to be the de facto standard for scientific and technical mathematical writing.

The issue with these projects is the presumptuous simultaneous breakage of the standard with a poor handpicked mangled subset thereof.

[1] https://ctan.org/tex-archive/macros/latex/required/amsmath

[2] https://docs.mathjax.org/en/latest/input/tex/macros/index.ht...

[3] https://katex.org/docs/supported.html

[4] https://typst.app/docs/reference/math/

vitorsron The microtype package: Subliminal refinements towards typographical perfection [pdf]

I should warn you that the stretch that the package performs does not preserve the vertical stem widths. Therefore, especially at small point sizes, there are perceptible weight changes.

In my opinion, the default font expansion value stretch=20 is not conservative enough. I would personally recommend not only reducing it but also using the selected argument to further reduce the expansion of glyphs whose geometry distortion may be more easily perceived.

vitorsron The microtype package: Subliminal refinements towards typographical perfection [pdf]

@dang, this URL submission points to a CTAN mirror, can you please change it to the following:

https://mirrors.ctan.org/macros/latex/contrib/microtype/micr...

vitorsron Ask HN: What's the state of the art for drawing math diagrams online?

Mermaid is unable to typeset math.

vitorsron Ask HN: What's the state of the art for drawing math diagrams online?

Just plain old draw.io:

https://www.drawio.com/doc/faq/math-typesetting

vitorsron Fourteen Years of Go

As I understand it, Go is funded by Google in the sense that core language development is largely undertaken by Google employees, not in the sense that an independent language foundation organization that receives funding from external backers invests in core language development.

vitorsron What's New in Python 3.12

I find the feature-happy kitchen sink product philosophy in programming languages quite insufferable.

It has become an unpleasant chore to keep up with PEP squabbles and accepted changes.

vitorsron A poor entropy coding result in Zstandard

I would not write this as failure of zstd.

There are no comments on particular command line usage of respective compression algorithm frontends.

zstd has higher compression levels that near LZMA and you can additionally train a specific dictionary to be used in compressing particular kind files.

vitorsron JPEG XL against AVIF tested on ImageEngine

The worst aspect of JXL is the chroma bleed:

https://afontenot.github.io/image-formats-comparison/#catedr... (glasswork)

https://afontenot.github.io/image-formats-comparison/#end-of... (teeth, fingernails)

https://afontenot.github.io/image-formats-comparison/#steinw... (lights, pianos)

https://afontenot.github.io/image-formats-comparison/#us-ope... (racket, fingernails, jewelry, watch)

vitorsron Modern Image Processing Algorithms Implementation in C

How does it compare to dlib [1], ccv [2] etc.?

[1] https://github.com/davisking/dlib

[2] https://github.com/liuliu/ccv

vitorsron Google open-sources Rust crate audits

See also Google Cloud’s Assured Open Source Software service:

https://cloud.google.com/assured-open-source-software

vitorsron Ask HN: What is new in algorithms and data structures these days?

- Streaming algorithms [1]

- Eventual consistency [2]

- Refinement types [3], [4]

[1] https://www2.cs.uh.edu/~ceick/6340/Streams.pdf

[2] https://www.microsoft.com/en-us/research/wp-content/uploads/...

[3] https://ranjitjhala.github.io/static/trust_but_verify.pdf

[4] https://ranjitjhala.github.io/static/refinement_types_for_ty...

vitorsron Companies hide negative news with unrelated press releases alongside SEC filings

Waiting for the Matt Levine column ‘Are Unrelated Press Releases Securities Fraud?’

vitorsron Tech CEO applauds an employee selling their dog to accommodate return-to-office

> He told content writers they should be using AI to increase output “30-50 times our normal production.”

Take it easy on the Kool-Aid.

vitorsron Show HN: Want something better than k-means? Try BanditPAM

Thanks for doing this, I will be sure to try it in ML competitions. I really like that you used Armadillo, which is something I personally want to do for my own projects.

Just out of technical curiosity, have you come across any particular developments or empirical evidence on the use of (invertible) data transformations to enhance clustering results? I am currently researching a particular problem within signal processing related to signal distribution transforms and I am particularly interested in reading about potential applications. As an example, and since you mention JPM partial funding, how would copula transformation affect the results of clustering (assuming an inverse exists etc. and we apply the inverse transformation afterwards)?

vitorsron FDIC Takes over Silicon Valley Bank

I have access to Refinitiv. I think it looks fine?

https://workspace.refinitiv.com/web/cms/?pageId=mbshome

vitorsron Raising the bar for software security: GitHub 2FA begins March 13

The average phone insurance where I live costs a 20% premium per annum relative to the value of the device due to the frequency phones are stolen.

What happens when my phone is eventually stolen? I must now depend on the goodwill of morose carriers?

vitorsron Even the Pylint codebase uses Ruff

I am happy if this tool works for you but I think writing off my comment as naysaying is not constructive.

There is no support for all syntax features of the current language specification and reference implementation. There is also no legacy support. Again, especially in projects concerned with code quality and analysis, building is not more important than long-term maintenance: Both language specification and reference implementation are constantly evolving. The current approach relies on an upstream interpreter and thereby inherits idiosyncratic divergences and limitations.

There is no extensibility. The majority of projects are neither public nor frameworks. Often projects employ a number of frameworks each with associated dialects, with both frameworks and dialects also constantly evolving. The current approach to extensibility is approximately one person manually rewriting rules individually from heterogeneous projects that span several contributors over several years. There is already outstanding maintenance work on manually introduced rules, only a subset of which gets flagged by a fraction of adopters. There is even outstanding work on feature parity.

vitorsron Even the Pylint codebase uses Ruff

To add to your comment:

I also think that these rewrites are throwing the baby out with the bathwater. Especially in projects related to code quality and analysis, things like widespread adoption and issues going back almost a decade are much more important than raw performance.

For every aspect of functionality, there is a heavy baggage of quirks, bugs, edge cases etc. I am not convinced that a rewrite of a tool that is still in development and still limited to a subset of functionality is able to capture nearly a decade of community work.

vitorsron Daniel Stenberg and the home of curl in Stockholm, Sweden

You can read it in the words of Daniel Stenberg himself:

https://un.curl.dev/money

vitorsron Only Debate The Non-Linear

I find that the kind of rhetoric espoused by the author is too eagerly neglectful of negative nonlinearities. An organization needs to survive before it can consider "optimizing for the first derivative".

vitorsron Scintilla is a free source code editing component with a permissive license

The Qt port of Scintilla is also used by a number of projects as a lightweight editor component:

  $ lsb_release -d && apt rdepends libqscintilla2-qt5-15
  Description:    Ubuntu 22.04.1 LTS
  libqscintilla2-qt5-15
  Reverse Depends:
    Depends: libqscintilla2-qt5-dev (<< 2.11.6+dfsg+1~)
    Depends: sqlitebrowser (>= 2.11.2)
    Depends: sonic-pi (>= 2.11.2)
    Depends: qgis-providers (>= 2.11.2)
    Depends: python3-qgis (>= 2.11.2)
    Depends: python3-pyqt5.qsci (>= 2.11.2)
    Depends: openscad (>= 2.11.2)
    Depends: octave (>= 2.11.2)
    Suggests: libqscintilla2-qt5-l10n
    Depends: libqscintilla2-qt5-dev (>= 2.11.6+dfsg)
    Depends: juffed (>= 2.11.2)
    Depends: libqscintilla2-qt5-designer (>= 2.11.2)
    Depends: libqgis-gui3.22.4 (>= 2.11.2)
    Depends: libqgis-app3.22.4 (>= 2.11.2)

vitorsron Google Python Style Guide

See also how yapf defines the Google formatting style:

https://github.com/google/yapf/blob/v0.32.0/yapf/yapflib/sty...

vitorsron Erlang: The coding language that finance forgot (2022)

I cannot name a single high-quality scientific computing and numerical algorithms Rust library that is not a thin layer over established C and C++ libraries.

vitorsron My Journey from R to Julia

    R
    Cited in: 8,589 Publications
    7,353 [Citing Publications in] Statistics (62-XX)
    https://zbmath.org/software/771

    Julia
    Cited in: 442 Publications
    64 [Citing Publications in] Statistics (62-XX)
    https://zbmath.org/software/13986

vitorsron My Journey from R to Julia

This has been said before multiple times over but with these languages it is rarely about the languages themselves but their ecosystems:

https://cran.r-project.org/web/packages/available_packages_b...

To go from R to Julia, as an example, one would have to give up on a hundred or so high-quality packages potentially related to their activities.

vitorsron I've procrastinated working on my thesis for more than a year

Some treacherous advice in the comments. I personally do not consider English literature and journalistic style guides to be applicable to primary scientific and technical literature writing without a diligent amount of scrutiny.

First and foremost, valid work needs to follow basic logical and empirical tenets. This is not easy nor does it come for free, and it is expected from primary scientific literature writers to take their time (or else risk engaging in low-quality work - or worse, pseudoscience).

Second, impactful work needs to easily connect to directly and indirectly adjacent literature. Truly impactful work needs to be relevant, correct, precise but understandable, comparable, reproducible and extendable, all in connection to related work. This is not easy nor does it come for free, and it is expected from writers to constantly check literature at every step before continuing (or else risk building on fragile ground - or worse, doing something that already exists).

Last, on doing something that already exists, novel work needs to not reinvent the many systems and subsystems it encompasses. Truly novel work needs a humble understanding of unknown unknowns. This comes at a hefty price of extensive learning and reading which is neither easy nor free.

Writing should be deferred to the consolidation of each incremental step related to primary scientific work.

Not writing should not be understood as procrastination of the scientific endeavor unless the work does not concern easing or putting you closer to steps' objectives.

My personal advice:

It is fine to reach blockers in scientific work. Other people certainly have reached similar blockers and not given up. Other people have also reached blockers and proposed alternative approaches. Work on: structuring ways to attack problems; attacking simpler or scaled-down versions of problems and formally classifying each version; mapping all different ways simpler problems can be modelled and attacked; strategizing how to reintroduce simplifications or scale; building solid foundations, sharing subproblems with colleagues or deferring certain knowledges outside the scope to the literature.

It is also fine to consider other lines of research if blockers are indeed unmanageable. Understand how the problem can be applied in other settings and consider pivoting to one particular setting. Understand as well that scientific work is not personal and you should not take inabilities as personal defeats.

Contributing knowledge to the scientific literature is hard, tiresome, and takes time and patience.

You want to run the distance, not sprint. You also want to make it to the end, not compete.

vitorsron Typst: A Programmable Markup Language for Typesetting [pdf]

Thanks. Yes, rendering and shaping are distinct but some of the linked libraries did one, the other, or both and the parent commenter singled out rastering which is how I ended up putting FreeType and HarfBuzz in the same sentence. Even then both are commonly used in tandem (see [1]-[9]) and have a few overlapping functionalities.

> it does support BiDi, complex script shaping

Hey, that is indeed quite good. Would you mind if I ask you how well is the support for popular Asian languages?

> linking C and Rust in WASM is unfortunately not really possible

Damn. I am not very experienced in Rust but I would not have guessed that. I apologize if I misrepresented difficulties related to targeting WASM.

[1] https://github.com/apple-oss-distributions/WebKit/tree/WebKi...

[2] https://github.com/apple-oss-distributions/WebKit/tree/WebKi...

[3] https://github.com/chromium/chromium/tree/main/third_party/f...

[4] https://github.com/chromium/chromium/tree/main/third_party/h...

[5] https://searchfox.org/mozilla-central/source/modules/freetyp...

[6] https://searchfox.org/mozilla-central/source/gfx/harfbuzz

[7] https://cs.android.com/android/platform/superproject/+/maste...

[8] https://cs.android.com/android/platform/superproject/+/maste...

[9] https://www.amazon.com/gp/help/customer/display.html?nodeId=...

vitorsron Typst: A Programmable Markup Language for Typesetting [pdf]

The linked libraries are not even close to solving limited subsets of problems solved by FreeType or HarfBuzz. No test is needed if they do not even have a working implementation of particular requisites: Do they work on heterogeneous layouts, directions, languages, locales, scripts, symbols and composites, extensions, variations, legacy, missing, partial or corrupted instructions, standards interpretations, platforms, output devices, nonstandard point structures and grids?

They do not. What they solve is almost a toy problem compared to the size, scope and breadth of these libraries.

Just because some project is implemented in Rust does not make it comparable never mind superior by default.

There is a world out there and it is not homogeneous format and standards-compliant Latin fonts in English LTR text in linear disposition with some generic rectangular subpixel rendering on a regular rectangular grid.

I warmly welcome you to browse closed issues of FreeType [1] and also the closed issues of HarfBuzz [2]. If you feel inspired please do also look into mailing lists and discussion pages related to the development, building, tracking and patching of packages of these projects in any of the numerous places it is used.

The only argument Rust people have is in relation WASM but if you insist in targeting WASM why not fork FreeType, strip it to the strict subset of features your application needs and target it?

Why do it in the first place? Why reinvent the wheel?

As such I will restate my view: I see no gain in using any of these subpar libraries.

[1] https://gitlab.freedesktop.org/freetype/freetype/-/issues/?s...

[2] https://github.com/harfbuzz/harfbuzz/issues?q=is%3Aclosed